Estimating Ocean Tides in the South China Sea by Assimilating Satellite Data Using 2D Tidal Adjoint Model and Equidistant Nodes Orthogonal Polynomial Fitting

Abstract The equidistant nodes orthogonal polynomial fitting (ENOPF) method proves effective and efficient to extract tidal information from satellite data, yet its dependency on accurate satellite data limits performance in coastal areas with sparse coverage, where extracting reliable harmonic constants is challenging. By using a two-dimensional (2D) adjoint tidal model and integrating X-TRACK data with simulated data via equidistant nodes orthogonal polynomial fitting, we extracted harmonic constant of four major tidal constituents (M 2 , S 2 , K 1 , and O 1 ) with high resolution in the South China Sea. We divided the study area into five subareas and mitigated systematic errors through varied weight coefficients for different data sources. For the M 2 and S 2 constituents, ENOPF had the smallest errors overall, while Finite Element Solutions 2014 (FES2014) and Empirical Ocean Tide 20 (EOT20) performed best for the K 1 and O 1 constituents. Although ENOPF was slightly less accurate overall, its coastal performance improved significantly with the 2D adjoint model. The cotidal charts from ENOPF were consistent with those from other models, though some distortion at the junction of area I and area V indicated a need for better splicing methods. Distortions in O 1 constituent simulations were likely due to strong internal tide signals. The findings provide a reliable high-resolution tidal dataset for the South China Sea, supporting accurate tidal corrections for sea level anomaly studies.